Classification of Forest Vegetation in North-Central Minnesota Using Landsat Multispectral Scanner and Thematic Mapper Data

Abstract

Abstract Computer classifications of Landsat-5 Thematic Mapper (TM) and Multispectral Scanner (MSS) data were evaluated to determine how forest and sensor characteristics affect the classification accuracy of Minnesota forest cover types. The test area was Itasca State Park in north central Minnesota. The experiments involved comparisons of sensors differing in spectral and spatial resolution. To evaluate classification performance the Landsat classification maps were compared on a pixel by pixel basis with a digitized reference map of the Park. Classification results were compared for statistically significant differences using discrete multivariate statistics. Classification accuracies ranged from 26 to 86%, depending upon the sensor, number of classes, and performance measure used. The most significant result was that the increased spectral/radiometric resolution of the TM data resulted in 7-15% absolute increase in forest cover classification accuracy over MSS data using conventional methods. The best spectral band combination was one band each from the visible, near infrared, and middle infrared. For. Sci. 36(2):330-342.